∮Research
Threads in progress.
A small number of long-running research threads, each building on a shared foundation: that perception, representation, and control are best understood as one continuous mathematical object.
- 01
Ontology Neural Networks
A learning architecture whose latent state carries the explicit type structure of a target ontology. Its original higher-order/cohomological thesis was audited to a scoped No-Go boundary; what survives is a modest direction-channel signal, a standard delay-margin certificate, and the audit methodology itself.
- 02
Perception theory
Soft Cognitive Cohesion — a mathematical theory of how coherent formations emerge prior to discrete objecthood. The graded field that precedes the world being parsed into separately-identified things. Aligned with Perception_theory canonical CV-1.17 (sealed 2026-05-15).
- 03
Mathematical foundations
RelationWorld Theory — a mathematically rigorous theory of discrete gauge structure on finite graphs, developing analogues of Yang–Mills, Uhlenbeck compactness, and Cheeger spectral geometry in a purely combinatorial setting.
- 04
Delay-robust control
Closed-loop control laws that remain stable under communication and sensing delay, built on top of an ontology-aware perceptual state. The ORTSF framework, reduced to its honest, verified content: a standard delay-margin certificate — decoupled from the ontology's cohomology, not a cohomological Lyapunov bound.
- 05
Robotics applications
Physical systems as test instruments for specific theoretical claims. The cognitive-robotics stack — SEGO architecture, semantic scene graphs, explainable control, and production-grade assembly on a Doosan M0609 — where the rest of the programme meets hardware.